panarchy and adaptive change - around the loop and back again

8/8/2019 Panarchy and Adaptive Change - Around the Loop and Back Again

1/21


2/21

Panarchy and Adaptive Change: Around the Loopand Back AgainBradley C. Karkkainen*

I. PANARCHY AND PROGRESSI am both honored and a bit humbled to comment on C.S."Buzz" Holling's contribution to this issue. Holling's influenceon environmental and natural resources law and policy overthe past several decades has been enormous. Holling pioneeredthe concept of "adaptive management"-or AdaptiveEnvironmental Assessment and Management (AEAM) as hefirst dubbed it in his seminal 1978 work on the subject'-as ascientifically defensible, learning-by-doing, incremental anditerative approach to improving our understanding of complex

ecosystems, leading to improvements in our ability to manageenvironmental problems in their proper ecological context.Holling developed the adaptive management concept in

response to his own frustrations as a scientist and frequentparticipant in interdisciplinary environmental impactassessment (EIA) teams in the 1970s. 2 The standard EIAapproach calls fo r a comprehensive and synoptic environmentalanalysis and assessment, legally required as a prerequisite toany governmental action that "significantly affects"environmental quality.3 Typically, however, only a fraction of

2005 Bradley C. Karkkainen.* Professor and Henry J. Fletcher Chair, University of Minnesota Law

School.1. See ADAPTIVE ENVIRONMENTAL ASSESSMENT AND MANAGEMENT (C.S.

Holling ed., 1978).2. Id. at 1-2.3. See National Environmental Policy Act 102(2)(C), 42 U.S.C.

4332(2)(C) (2000) (requiring federal agencies to "include in everyrecommendation or report on proposals for legislation and other major Federalactions significantly affecting the quality of the human environment, adetailed statement" on environmental impacts and alternatives to theproposed action).

HeinOnline -- 7 Minn. J.L. Sci. & Tech. 59 2005-2006


3/21

MINN. J.L. SCI.& TECH.the information needed to make such a comprehensiveassessment is readily available. Data gaps are pervasive,scientific understanding of even the most basic ecologicalprocesses is often incomplete, and the integrative analyses reston crude, untested, incomplete, and necessarily speculativeecosystem models stitched together from spotty data anddisparate scraps of research findings from a variety of scientificdisciplines. Worse, the validity of the predictions generated bythese models is rarely tested against real-world outcomes dueto a paucity of post-project monitoring. Consequently,scientists have little opportunity to refine or adjust theirmodels over time.

Holling believed such efforts led neither to better scientificunderstanding nor to well-informed agency decisionmaking.For their part, agency managers seemed to have little concernfo r the accuracy of the assessments generated by this jerry-rigged process, so long as the legally mandatory procedureswere followed. 4 As a legal matter, it is procedure and notsubstance that matters in environmental impact assessment inthe United States.5A more scientifically defensible approach, Holling argued,would be to treat environmental impact assessment not as aonce-off, purely predictive exercise, but rather as an ongoing,interdisciplinary, scientific inquiry.6 Teams of scientists wouldconstruct integrative models based upon the best data andresearch currently available, then identify gaps anduncertainties in their models and generate testable hypothesesdesigned to fill the gaps and reduce the uncertainties. Workingwith managers, the scientists would then design managementinterventions as scientific experiments carefully tailored to

4. See COUNCIL ON ENVTL. QUALITY, THE NATIONAL ENVIRONMENTALPOLICY ACT: A STUDY OF ITS EFFECTIVENESS AFTER TWENTY-FIvE YEARS, atiii (1997) (concluding that many agencies act "as if the detailed statementcalled for in the statute is an end in itself, rather than a tool to enhance andimprove decision-making").

5. See A. Dan Tarlock, Is There a There There in EnvironmentalLaw?, 19J. LAND USE & ENVTL. L. 213, 239 (2004) (describing NEPA as a "strikingexample of the dominance of procedure over substance," a condition traceableto early court decisions holding that only the statute's procedural provisionswere judicially enforceable).

6. See ADAPTIVE ENVIRONMENTAL ASSESSMENT AND MANAGEMENT,supra note 1, at 133 ("Prediction and traditional 'environmental impactassessments' suppose that there is a 'before and after,' whereas environmentalmanagement is an ongoing process .... Environmental assessment should bean ongoing investigation into, not a one-timepredictionof, impacts.").

[Vol. 7:1



4/21

PANARCHY AND ADAPTIVE CHANGEfield-test their hypotheses against observed outcomes. Theresults of these "adaptive management" experiments wouldfeed back into further refinements of the ecological models,generating subsequent rounds of testable hypotheses. Iterativeapplication of this method, Holling argued, would lead torolling improvements in scientific understanding andcontribute in turn to better informed and continuouslyimproving management decisions.7

It took the better part of two decades for the adaptivemanagement concept to catch on in natural resourcesmanagement, and we are yet to appreciate its full implications.Although many leading scientists and natural resourcesmanagement professionals now see the need for some form ofadaptive management as axiomatic,8 even the most well-fundedand technically sophisticated ecosystem management efforts,like those in the Florida Everglades and the San FranciscoBay-Delta, are still struggling, often awkwardly anduncertainly, to integrate adaptive management principles. 9

Environmental law and policy scholars are several steps

7. See id. at 11-16 (describing principal elements in an AdaptiveEnvironmental Assessment and Management approach).8. See Fred Bosselman, A Role for State Planning: IntergenerationalEquity and Adaptive Management, 12 U. FLA. J.L. & PUB. POL'Y 311, 323-25

(2001) (describing rising prominence of adaptive management in ecologicalscience); A. Dan Tarlock, PuttingRivers Back in the Landscape:The Revival ofWatershedManagement in the United States, 6 HASTINGS W.-Nw. J. ENVTL. L.& POL'Y 167, 189-92 (2002) (describing the shift in natural resourcesmanagement toward regional ecosystem-scale efforts at integratedmanagement, predicated upon adaptive management strategies).

9. See NATIONAL RESEARCH COUNCIL, COMM. ON RESTORATION OF THEGREATER EVERGLADES ECOSYSTEM, ADAPTIVE MONITORING AND ASSESSMENTFOR TH E COMPREHENSIVE EVERGLADES RESTORATION PLAN 9 (2003)(concluding that monitoring, assessment, and passive adaptive managementplans for the Comprehensive Everglades Restoration Project are scientificallydefensible, but should be augmented with active adaptive management andmeasures to ensure that information reaches key decisionmakers); Thomas T.Ankerson & Richard Hamman, Ecosystem Management and the Everglades:ALegal and InstitutionalAnalysis, 11 J. LAND USE & ENVTL. L. 473, 500 (1996)(concluding that "the implementation of adaptive management policies isproblematic [in the Everglades] under the current environmental regulatoryframework" which is predicated upon a static approach to environmentalmanagement); J.B. Ruhl, Taking Adaptive Management Seriously: A CaseStudy of the Endangered Species Act, 52 U. KANSAS L. REV. 1249, 1265-67(2004) (stating that despite its professed commitment to an adaptivemanagement approach, the CALFED Bay-Delta program recentlyacknowledged that it had failed to conduct regular status monitoring on theendangered Delta smelt, a necessary predicate for meaningful adaptivemanagement).

2005]



5/21

MINN. J.L. SCI. & TECH.further behind the curve. We are only now beginning to debatethe meaning and merits of adaptive management in a seriousway and to appreciate the profound and far-reaching challengeit poses to more familiar fragmentary, fixed-rule approaches tolaw and regulation. J.B. Ruhl's contribution to this issue is animportant intervention in that debate,10 which I discuss below.Now Buzz Holling has come forth with a new and equallyrevolutionary concept, "panarchy." I confess that at first blushI am not entirely certain what to make of it, and suspect I amnot alone in that regard. Perhaps we legal scholars are just abit slow on the uptake, but Holling appears once again to beway out ahead of us, setting his sights on some distant anddimly perceived horizon of knowledge. And as with adaptivemanagement, it may take us another twenty years to catch up.Holling, if you will pardon the analogy, is the Lance Armstrongof the adaptive change cycle; we feel privileged just to besomewhere far back in the pack chasing him, however futilemay be the effort to catch him.

"Panarchy," as Holling and his colleagues describe it, is acycle of adaptive change, proceeding through "forward-loop"stages of innovation, growth, exploitation, consolidation,predictability, and conservation, followed by "back-loop" phasesof instability, release, collapse, experimentation, novelrecombination, and reorganization.1 At that point the cyclebegins anew, moving once again through the forward loop,albeit beginning from a new starting point. Holling and friendsschematize this cycle by simplifying it to four phases:exploitation, conservation, release, and reorganization. 12 Thiscycle, they argue, generally characterizes processes of adaptivechange in both ecological and social systems. 13So far, so good. From there, though, the story becomesvexingly complex. These adaptive cycles, Holling tells us,operate simultaneously on multiple spatial and temporalscales, and across scales. Consequently, "major learned

10. J.B. Ruhl, Regulation by Adaptive Management-Is It Possible?, 7MINN. J.L. Sci.& TECH. 21 (2005).

11. See C.S. Holling & Lance H. Gunderson, Resilience and AdaptiveCycles, in PANARCHY: UNDERSTANDING TRANSFORMATIONS IN HUMAN ANDNATURAL SYSTEMS 25, 33-40, 47-49 (Lance H. Gunderson & C.S. Holling eds.,2002) [hereinafter PANARCHY]; C.S. Holling, From Complex Regions toComplex Worlds, 7 MINN. J.L. Sci. & TECH. 1 (2005) [hereinafter ComplexRegions].

12. Holling & Gunderson, supra note 11 , at 34 fig.2-1.13. See id. at 62.

[Vol. 7:1



6/21

PANARCHY AND ADAPTIVE CHANGEbenefits need not, and generally do not stay in the place wherethey were created," but instead "flourish elsewhere." 14 Thussuccessful local-scale innovations might create newopportunities that are absorbed into larger forward-looplearning processes, and become diffused across larger spatialand temporal scales. By the same token, collapse at smallscales might contribute to cascades of destabilization andcollapse across larger scales. Or, locally adaptive successesmight in some circumstances prove disruptive of larger scalesof forward-loop self-organization, triggering a flip into back-loop instability and collapse at the larger scale. 15

Holling and his colleagues claim to have documented therecurrence of these cyclical patterns in numerous ecologicalsettings, 16 and they cite a variety of social science literatureand anecdotal accounts to argue that the pattern is generallyreplicated in human social systems as well. 17

Ultimately, of course, the panarchy thesis rests on a seriesof large empirical claims. Those claims, while plausible andintriguing, are difficult to verify or to rebut from the vantagepoint of a law professor's desktop, and therefore I have nochoice but to remain agnostic-at least for now. A fewcomments are nonetheless in order.

The panarchy thesis, as I understand it, is posited as ageneral characteristic of cycles of adaptive change. This lendsa note of determinism and inevitability, something like an IronRule of the Figure-8 (or "double-loop") pattern of systemchange. Holling and colleagues are quick to disclaim a stronglydeterministic interpretation of their own thesis, labeling it ametaphor" rather than a "rigid, predetermined path."18 But a

14. See Complex Regions, supra note 11 , at 4.15 . See Holling & Gunderson, supra note 11 , at 60-61 (arguing that locally

adaptive strategies may sometimes have maladaptive large-scaleconsequences, as when efforts to stabilize resource production reduce diversityand resilience, leading to collapse in the larger system).

16. See id. at 35-38 & box 2-2 (describing this adaptive change cycle inseveral ecosystem contexts).17. Id. at 38, 55-60 & box 2-5 (describing adaptive change cycle in

corporate management); Fikret Berkes & Carl Folke, Back to the Future:Ecosystem Dynamics and Local Knowledge, in PANARCHY, supra note 11, at121-46 (describing panarchic adaptive changes cycles in local and traditionalnatural resources management regimes).

18. See Holling & Gunderson, supra note 11, at 51 ("The adaptive cycle inits most general form is a metaphor and should not be read as a rigid,predetermined path and trajectory-for ecosystems at least, let aloneeconomies and organizations.").

2005]



7/21

MINN. J.L. SCI. & TECH.metaphor for what? Presumably, it is a metaphor for anobserved phenomenon that, according to their thesis, recurswith predictable regularity. Holling's own writings elsewheresuggest that the exact course and direction that change takesin any particular complex dynamic system will necessarily beunpredictable, nonlinear, and stochastic, 19 yet the panarchythesis seems to imply that change, however unpredictable inthe particulars, will always (or almost always) follow a familiarand predictable progression of steps. The pattern, and thepattern alone, is presented as certain and more or lessinescapable. Why that should be is never fully explained. Tothat extent, although highly abstract, the panarchy thesisseems a bit undertheorized.

Next, it is not entirely clear what to make of the role ofhuman agency in a panarchic world, especially as it applies tohuman social systems. Certainly we humans play a pivotal rolein social systems-and for that matter, in most ecologicalsystems, too, given our power to dominate and displace otherspecies and to disrupt or alter ecological processes. 20Sometimes our impact stems from, or at least is modified by,conscious, intelligence-informed acts. Presumably, then, ourunderstanding of the dynamics of system change itself has thepotential to change the way we act within the system, therebyaltering the trajectory of system change-though perhaps,given complexity, nonlinearity, and inherent stochasticity, inways we cannot entirely predict or control. Holling andcompany do not tell us much about this, sometimes writing as ifit is simply all beyond our ability to comprehend and manage,and we are all just along for the ride on the "Double LoopPanarchy Express."

19. See C.S. Holling, Fikret Berkes & Carl Folke, Science, Sustainability,and Resource Management, in LINKING SOCIAL AND ECOLOGICAL SYSTEMS:MANAGEMENT PRACTICES AND SOCIAL MECHANISMS FOR BUILDINGRESILIENCE 342, 352 (Fikret Berkes et al. eds., 1998) (stating that naturalresources management problems exhibit "aspects of behavior [that] arecomplex and unpredictable and . . .causes, while at times simple (when wefinally understood), are always multiple"); see also id. at 346-47 (stating thatin complex natural systems "uncertainty is high," "knowledge of the system wedeal with is always incomplete," "[s]urprise is inevitable," and "the systemitself is a moving target").

20. See Frances Westley et al., Why Systems of People and NatureAre NotJust Social and Ecological Systems, in PANARCHY, supra note 11 , at 104(describing competing theories of human domination of ecosystems, allpredicated upon our unusually large capacity to change or destroy the naturalenvironment).

[Vol. 7:1



8/21

PANARCHY AND ADAPTIVE CHANGEThe implications seem pretty scary. It is a bit like beingtrapped on a twenty-first century, high-tech version of a

double-looping carnival ride, armed with the capacity to fiddlewith the precise trajectory but only in ultimately unpredictableways, and never able to change the basic pattern of ourinterminable trip: forward, up, around, down, back loop, down,around, up, forward again, and the more we try to adjustcourse, the more the pattern stays the same. It sounds like amad ecologist's vision of a living hell, a Huis c1os2 1 of adaptivechange.

Holling's earlier work on adaptive management left moreroom for optimism. It gave us affirmative reasons to embraceepistemic humility. Recognition that ecological systems aresimply too complex to fully comprehend may have seemeddebilitating at first, given a culture of regulation and resourcemanagement that assumed it was the regulator's or manager'sresponsibility to have all the scientifically "correct" answersbefore proceeding. Complexity thus initially threatened tothwart action. Adaptive management seemed to offer a wayout of that box: an incremental, scientifically defensible, one-step-at-a-time methodology for learning as we go, and adjustingmanagement measures to incorporate gains in knowledge overtime.

Panarchy, on the other hand, conjures up not only intricatelayers of impenetrable and seemingly unmanageablecomplexity, but also a sense of futility, coupled with less clarityand optimism about how to respond. The panarchy thesisseems to imply that all directed, "forward-loop," human socialendeavors-adaptive ecosystem management included-aredoomed to fail eventually, perhaps after an initial period ofsuccess. Despite our best efforts, the cycle from innovationthrough growth, exploitation, maturity, stability, andconservation appears destined inevitably to stumble into theback-loop of destabilization, disorder, and creative destruction.Every project, large or small, thus takes on an aspect of thelabor of Sisyphus-or Bill Murray's Groundhog Day withoutthe happy ending, 22 if your tastes run more to pop culture-but

21. The reference is to Jean-Paul Sartre's existentialist play Huis clos(translated as "No Exit") in which three characters are condemned to spendeternity sitting on a couch conversing, even though they cannot stand eachother.

22. GroundhogDay, a 1993 film directed by Harold Ramis, starred comicactor Bill Murray as a television weathercaster doomed to spend eternity in

2005]



9/21

MINN. J.L. SCI. & TECH.either way, it is a gloomy prospect. If indeed all humanundertakings are doomed to fail at unexpected times and forunanticipated reasons, why bother? Unless, of course, we areso biologically programmed that we have no other choice, whichis an equally dismal thought.

But what about the "back loop" of the panarchic cycle?That idea has great intuitive appeal. We have all seen cycles ofcreative destruction, alternating with periods of incrementalfront-loop learning. The front loop imposes order and stability(at least for a while), but at the same time grows increasinglyvulnerable to disturbance until suddenly everything unravelsin a flurry of rapid, back-loop, creative disorganization,reorganization, and reconfiguration, leading to innovation, re-synthesis, and a new front-loop developmental pattern. It is onthe back loop that the most creative and far-reaching changesare often inaugurated in the midst of seeming chaos anddisintegration. This is no less true in law than in other fields ofhuman endeavor. It took the American Revolution and asubsequent period of futility under the Articles ofConfederation (back loop) to spawn a long period of relativestability under the United States Constitution (front loop); theCivil War (back loop) to generate the ReconstructionAmendments, a long period of slowly accumulating federalconstitutional primacy, and steady advances in civil rights andcivil liberties (front loop); the Great Depression (back loop) toproduce the New Deal and the modern administrative andregulatory state (front loop).

In the environmental arena, too, it is typically periods ofcrisis that stimulate creative and innovative advances in policyand management. Thus it took the Santa Barbara Channel oilspill and the Cuyahoga River bursting into flames (back loop)to bring about the Clean Water Act (front loop),2 3 Love Canal toproduce CERCLA, 24 Bhopal to launch the Toxics ReleaseInventory, 25 and the Exxon Valdez to spur enactment of the OilPollution Act. 26 Collaborative, adaptive, and integrative

endless repetition of the worst day of his life, covering the annual emergenceof the groundhog in a backwater town-until the inevitable happy ending.

23. Michael P. Vandenbergh, The Social Meaning of EnvironmentalCommand and Control, 20 VA. ENVTL. L.J. 191, 206 (2001).

24. Id. at 206 n.83.25. See Christopher H. Schroeder, Third Way Environmentalism, 48 U.

KAN. L. REV. 801, 818 (2000).26. See Vandenbergh, supra note 23, at 206 n.83.

[Vol. 7:1



10/21

PANARCHY AND ADAPTIVE CHANGEenvironmental governance institutions have emerged in placeslike the Florida Everglades only when it became clear thatincremental business-as-usual was not working, the centerwould not hold, and something radically new and differentneeded to be tried.2 7 In short, we appear to be most open toexperimentation and new thinking when we are desperate andno longer can see the way to answers through accustomedmodes of thought and familiar patterns of action. Perhaps thatis where, at the end of the day, a glimmer of hope is to be foundin panarchy.

My friends and sometime collaborators, Chuck Sabel andBill Simon, have written provocatively about a legal approachthey call "destabilization rights."2 8 In Sabel and Simon's view,sometimes it becomes necessary to use litigation to pull theplug on an established institutional arrangement when itbecomes clear the institution is failing on one or moreimportant dimensions of performance, and political blockage ispreventing a significant correction through ordinarygovernmental processes. 2 9 Increasingly, Sabel and Simonargue, judges in "public law litigation" cases-typicallyconstitutional or statutory civil rights or civil liberties casesseeking major institutional reforms of public schools, prisons,criminal justice systems, social welfare programs, and thelike-recognize that they are not well-positioned to imposecomprehensive institutional reform blueprints from the benchgiven the limitations of judicial competence and expertisecoupled with limitations on the courts' ability to craft andeffectively enforce a new institutional order on a resistant,entrenched, governmental defendant. 30 So instead judges seek

27. See generally Alfred R. Light, Ecosystem Restoration in the Everglades,14 NAT. RESOURCES & ENV'T 166 (2000) (describing how an ambitious federal-state collaborative Everglades ecosystem restoration plan emerged out of thesettlement of an acrimonious lawsuit over water quality standards which hadtaxed the resources of both state and federal governments without producingimprovements in environmental quality).

28. See generally Charles F. Sabel & William H. Simon, DestabilizationRights: How Public Law LitigationSucceeds, 117 HARV. . REV. 1015 (2004).

29. Id. at 1062 (stating that the "prima facie case for public lawdestabilization has two elements: failure to meet standards and politicalblockage").

30. Id. at 1053 ("Courts found they lacked both the information and thedepth and range of control to properly formulate and enforce command-and-control injunctions" which "exacerbated resistance on the part of defendants -or at least, top-down measures had little capacity to neutralize suchresistance.").

2005]



11/21

MINN. J.L. SCI. & TECH.

to disentrench and destabilize the offending institution,ordering its disestablishment and setting out only very generalperformance standards for its reconstituted successor, leavingthe hard work of institutional redesign to the defendantgovernment but retaining jurisdiction to review its progressand to order additional reforms if necessary.3 1 This judiciallyimposed destabilization, Sabel and Simon argue, often opensthe door to creative, multiparty collaborations in theinstitutional redesign and reform effort, leading to a new andimproved institutional synthesis. In short, Sabel and Simon's"destabilization rights" look like a judicially triggeredmechanism to get to panarchy's back loop, where creativitymay flourish.

Citizen suits can sometimes perform a similar function inenvironmental law. Litigation-or sometimes the merebackground threat of litigation-has created legal, political,and institutional crises that have forced parties to cometogether in new and sometimes surprising reconfigurations inthe Florida Everglades, the San Francisco Bay-Delta, thesouthern California coastal sage scrub, and the PacificNorthwest, among others. In all these cases, legaldestabilization of established ways of doing business openedthe door to a new back-loop period of creative ferment,reorganization, and institutional innovation. The panarchythesis suggests that the new institutional configurations thatemerge out of this process can be expected to advance in theirturn through a normal front-loop cycle of growth, consolidation,and predictability, eventually to become unstable and collapseat unexpected times and in unanticipated ways into new backloops of creative destruction.

These examples suggest an intriguing prospect: suppose weintentionally seek to build back-loop cycles of destabilizationand back-loop reconfiguration into larger front-loop cycles ofinstitutional growth and consolidation? Then the inevitabledestabilization and collapse of front-loop processes need notlook so frightening or debilitating. From that perspective, thepanarchic double-loop comes to represent not the futile labor ofSisyphus, but a recurring cyclical pattern of alternatingopportunities for back-loop creativity and front-loopconsolidation-and maybe, just maybe, the opportunity for

31. Id. at 1067-73 (describing the general outlines of experimentalistremedies in destabilization rights cases).

[Vol. 7:1



12/21

PANARCHY AND ADAPTIVE CHANGEprogress over time, as each turn around the double-loopincorporates and builds on learning from past trips.

Moreover, the theory also suggests that it may be possibleto build smaller, controlled cycles of destabilization and back-loop creative reconfiguration into larger scale forward-loopprocesses, building toward greater order at the larger scalewhile learning as we go through a series of small-scaledisturbances and innovations. That, in a sense, is whatadaptive management-as Holling originally conceived it-isall about: creating small, controlled perturbations in the systemthrough carefully crafted, hypothesis-testing managementinterventions, and incorporating the learning thus generatedinto forward-loop incremental improvements in management atthe larger system-wide scale. Yet surely there will be large-scale surprises as well, leading at times to destabilizationacross the larger-scale system. The lesson of panarchy,however, is not to worry: change is inevitable, but back-loopdestabilization is both an inevitable part of the cycle of change,as well as an extraordinary learning opportunity, a time whenconditions are ripest for creativity to flourish and for learningto advance in great strides through creative recombination.Maybe this carnival ride will be fun after all.

II. ADAPTING LAW TO ADAPTIVE MANAGEMENTThat brings me to J.B. Ruhl's thoughtful and important

contribution, entitled Regulation by Adaptive Management-IsIt Possible?.32 Ruhl argues persuasively that adaptivemanagement is fundamentally incompatible with our presentparadigm of administrative law. It is only at the end of thepaper, however, that Ruhl finally takes a stand in favor ofmodifying administrative law to accommodate adaptivemanagement, rather than simply concluding that adaptivemanagement is impossible in practice-a conclusion to whichmany administrative law scholars come to far too easily, as ifadministrative law were somehow immutable and eternal, or atleast of constitutional stature, rather than just anotherstatutory and judge-made legal artifact that may provemaladaptive at some point.

I have no basic disagreement with Ruhl's thesis, but I dohave a couple of comments. First, I would urge that we paycloser attention to the forms and varieties of adaptive

32. Ruhl, supra note 10.

2005]



13/21

MINN. J.L. SCI. & TECH.management. A common theme in the scientific literature onadaptive management is the distinction between its "passive"and "active" forms. Carl Walters, a noted fisheries biologist,adaptive management theorist and practitioner, and frequentcollaborator with Buzz Holling, defines "active" adaptivemanagement as a "deliberate probing for information" througha multistep process involving integrative ecological modeling,conscious generation of testable scientific hypotheses, and fieldexperimentation through carefully tailored managementinterventions designed to test specific hypotheses. 33 "Active"adaptive management is, then, as much about harnessingmanagement interventions in the pursuit of science as it isabout harnessing advancing knowledge in the interest of soundmanagement. 34

In contrast, "passive" adaptive management is a simplerprocess involving heightened monitoring of key indicators,leading to subsequent adjustments in policies in light of whatmay be learned through careful observation and datageneration, and without the "deliberate probing" of active,hypothesis-testing experimentation. 35 Even "passive" adaptivemanagement is, in Walter's view, far superior to old-fashioned"trial-and-error" approaches in which the manager simply triesout a policy thought likely to succeed and then abandons it infavor of an alternative course of action if the first attemptfails. 36 Clearly, however, Walters sees a hierarchy of methods,preferring "active" to "passive" adaptive management whereverthe former is possible.

It is the "passive" form of adaptive management thatappears to have gained the greatest foothold in naturalresources management, however, and that is the form to whichJ.B. Ruhl refers in his paper. Ruhl offers his own rough-and-ready definition. The "essence" of adaptive management, hesays, is "an iterative, incremental decisionmaking process builtaround a continuous process of monitoring the effects ofdecisions and adjusting decisions accordingly."3 7 Thatdefinition comes very close to capturing Walters' notion of

33. CARL WALTERS, ADAPTIVE MANAGEMENT OF RENEWABLE RESOURCES232 (1986).

34. See id. at 64, 232.35. See id. at 232, 248-52.36. See id. at 64 (characterizing "trial-and-error" as a process of "blindprobing" without model-building and hypothesis-testing).37. Ruhl, supranote 10, at 28.

[Vol. 7:1



14/21

PANARCHY AND ADAPTIVE CHANGE

"passive" adaptive management, turning as it does onmonitoring and subsequent adjustment of policies. Ruhl alsocites definitions by biologist Simon Lewin and the NationalResearch Council that also track the "passive" rather than the"active" form of adaptive management. 38 None of theseformulations comes close to the highly structured andrigorously science-driven form of "active" adaptive managementadvanced by Walters and Holling as a disciplined way of field-testing specific and carefully formulated scientific hypotheses.

As applied in the real world of natural resourcesmanagement, however, what passes under the banner of"adaptive management" often does not even rise to the level of"passive" adaptive management as Walters defines it. Forexample, the United States Fish and Wildlife Service (FWS)-the federal agency in charge of most Endangered Species Actimplementation and enforcement-uses its own idiosyncraticdefinition of "adaptive management" within its HabitatConservation Plan (HCP) program. FWS defines adaptivemanagement as consisting of a series of pre-specifiedcontingency measures that will be adopted at pre-specifiedtriggering thresholds if the initial effort fails to produce theexpected results. 39 Thus, for example, an approved HCP mayspecify that if the population of a protected butterfly falls belowa specified level, the landowner will be required to spend anadditional thirty percent beyond sums already irrevocably

38 . See id. at 28 n.14 (citing Lewin's definition of adaptive management);id. (citing National Research Council description of adaptive management);see also id. at 29 (describing an eight-step process of adaptive management);id. at 34 (schematizing adaptive management as a four-stage continuous loopprocess). "Active" adaptive management as defined by Walters would includethe development of specific, testable scientific hypotheses and the design andimplementation of management measures carefully tailored to test thosehypotheses, steps mentioned nowhere in any of the definitions or descriptionof adaptive management cited by Ruhl.

39. See U.S. FISH & WILDLIFE SERV. & NATL MARINE FISHERIES SERV.,ENDANGERED SPECIES HABITAT CONSERVATION PLANNING HANDBOOK 3-25(1996) (stating that adaptive management plans in HCPs should specifytriggering thresholds and "a clear understanding and agreement" as to the'range of adjustments which might be required"); U.S. Fish & Wildlife Serv. &Nat'l Marine Fisheries Serv., Notice of Availability of a Final Addendum to theHandbook for Habitat Conservation Planning and Incidental Take PermittingProcess, 65 Fed. Reg. 35,242, 35,253 (June 1, 2000) (stating that if an HCP"incorporate[s] an adaptive management strategy, it should clearly state therange of possible operating conservation program adjustments due tosignificant new information, risk, or uncertainty," and that this range "definesthe limits of what resource commitments may be required of the permittee").

2005]



15/21

MINN. J.L. SCI. & TECH.committed for purposes of removing additional invasivevegetation and replanting with native vegetation. As onecommentator succinctly put it, "another word for adaptivemanagement" as practiced by the FWS "is 'contingencyplanning."'40

Now contingency planning may be a step up insophistication from old-fashioned management by "trial-and-error" (or "muddling through"), but it does not have the opentexture, flexibility, unboundedness, and openness to surpriseand unanticipated changes contemplated by advocates ofadaptive management. Not only does it lack the scientific rigorand experimental flavor of "active" adaptive management, butit does not even rise to the level of "passive" adaptivemanagement as described by Walters. It is probably not fairlycharacterized as "adaptive management" at all on moststandard definitions of that term. To use the term Ruhlborrows from Shapiro and Glicksman, 41 "contingency planning"is at heart just a slightly more complex form of "front-end"decisionmaking, modified only by a relatively narrow andinflexible range for "back-end" adjustments-withinparameters specified at the front end-when certain triggeringthresholds (also specified at the front end) are met.Contingency planning, in short, offers little opportunity forlearning and even less for rigorous scientific experimentation;ordinarily these are much more open-textured undertakings.

Be aware, then, that when FWS says "adaptivemanagement," it is using the term as an idiosyncratic term-of-art for ordinary contingency planning. When Buzz Holling andCarl Walters use the term "adaptive management," they meansomething very different and are generally referring to arigorously scientific form of "active" adaptive management,unless they specify that they are talking about the "passive"variety. When J.B. Ruhl and the National Research Counciluse the term, they are using it in yet another sense, to meanwhat Walters calls "passive" adaptive management. Thesevaried uses of the term invite confusion.

40. Gregory A. Thomas, Where PropertyRights and Biodiversity ConvergePart11: IncorporatingAdaptive Management and the PrecautionaryPrincipleinto HCPDesign, 18 ENDANGERED SPECIES UPDATE, 32, 33 (2001).

41. Ruhl, supra note 10, at 30 & n.18 (citing SIDNEY A. SHAPIRO &ROBERT L. GLICKSMAN, RISK REGULATION AT RISK: RESTORING A PRAGMATICAPPROACH (2003) for the distinction between "front-end" and "back-end"decisionmaking).

[Vol. 7:1



16/21


My point here is more than a mere syntactical one. Weneed to be careful how we use the term "adaptivemanagement." Participants in the debate over its meaning andmerits often end up talking past each other, because differentpeople are using the same term to mean very different things.A little more precision is advised.

Perhaps this explains, in part, the apparent paradox thatlies at the heart of FWS's HCP program, which professessimultaneously to embrace both "adaptive management" and a"no surprises" policy, the latter offering firm assurances tolandowners that more will not be demanded of them at the"back end" than they agree to accept at the "front end" of theplanning process. FWS insists these two principles are not inconflict. J.B. Ruhl agrees with that statement, arguing that ifthe landowner agrees at the front end to an HCP containing anadaptive management provision, that landowner cannot laterclaim at the back end to be "surprised" if the learninggenerated by adaptive management leads to a call for newactions-because that is precisely what the landowner agreedto.4 2 But Ruhl's conception of ("passive") adaptive managementis far more open-ended than FWS's notion of adaptive-management-as-contingency-planning. If "adaptivemanagement" means nothing more than a limited form ofcontingency planning, then adaptive management and nosurprises are obviously not in conflict, because the full range ofpossible outcomes, and the conditions that will trigger each ofthem, have already been fully specified at the front end. Thereis no conflict, but there is also no adaptive management in theopen-ended but "passive" sense advanced by J.B. Ruhl, muchless in the "active" scientific-experimentational sense advancedby Buzz Holling and Carl Walters. We are simply comparingapples, oranges, and pears and calling them all "apples."

I have no reason to question J.B. Ruhl's larger claim thatthe adversarial and litigious character of contemporaryadministrative law coupled with its overall tendency towardnitpicking enforcement of fixed "command-and-control" rules--especially procedural rules, which are singularly easy for courtsto enforce-and its reluctance to countenance uncertainty andlack of information as the basis for agency decisionmaking areall profoundly at odds with the very concept of adaptivemanagement. That strikes me as not only plausible, but very

42. See Ruhl, supra note 10 , at 47-48.

2005]



17/21

MINN. J.L. SCI. & TECH.probable, and it is a very serious problem because it places lawsquarely at odds with science, a very uncomfortable place forenvironmental law in particular to find itself.

I do question, however, whether HCPs are the case thatbest illustrates Ruhl's point. If I am right that FWS's crabbedand idiosyncratic interpretation of "adaptive management" as anarrowly circumscribed form of front-end contingency planningshould not count as true "adaptive management" at all (onstandard definitions of that term), then it appears that FWSmay never have really tried to incorporate genuine adaptivemanagement (as the rest of us know it) into the HCP process.In short, HCPs are not adaptive because FWS has never reallytried to make them adaptive, and not because environmentalplaintiffs have knocked out a few HCPs on administrative lawgrounds. It is premature to declare the policy experiment afailure, because we have not yet conducted the experiment-atleast, not in the HCP context.

That does nothing to make things easier for adaptivemanagement, however. It only leads me to conclude that, sofar, FWS has determined it must "talk the talk" by using thelanguage of adaptive management to try to justify its policies tothe scientific community and other informed observers, but ithas not yet "walked the walk" by actually implementing ameaningful form of adaptive management in practice. It isentirely possible, though I do not know for sure, that it is notthe presence of genuine adaptive management in the cases Ruhldescribes, but rather its absence, that led to lawsuits.

In the beach mouse case, for example, 43 the agency'sposition seems to have been that we genuinely do not knowwhat the impact of the proposed development on the protectedspecies will be, but we will promise only to make upward ordownward adjustments in mitigation measures within a pre-specified range, whether or not adjustments in that range willbe sufficient to protect the species. In such circumstances, alawsuit by environmentalists claiming that the agency is notentitled to reach a "finding of no significant impact" (FONSI)when it has in effect admitted that there could be a significantadverse impact on a species protected by the EndangeredSpecies Act seems entirely reasonable. If, on the other hand,the HCP had incorporated a more open-ended form of adaptivemanagement, promising to make whatever adjustments were

43. See Ruhl, supra note 10, at 51-52 nn.98-99.

[Vol. 7:1



18/21


necessary to maintain a viable population in light of what isrevealed by monitoring or scientific hypothesis testing, then theagency would have a more plausible claim to a FONSI, andenvironmentalists would presumably have less room forcomplaint, either on procedural grounds or with respect to thesubstantive outcome. In the agency's calculation, however, theNo Surprises principle trumps genuine adaptive management;No Surprises is compatible only with the faux adaptive-management-as-limited-contingency-planning that the agencyidiosyncratically advocates, and it is the application of thatpolicy that both inspires the environmentalists' distrust andinvites their lawsuit.

My second point is briefer. J.B. Ruhl argues that we arestill a long way from drafting a National Adaptive ManagementAct. 44 That is surely right, but merely raising the idea is apositive contribution. After all, adaptive management is atbottom a set of procedural principles-simultaneously a methodof inquiry and a procedural mechanism of agencydecisionmaking, based on rigorous observation throughmonitoring ("passive") and experimentation ("active"),reassessment, and adjustment in light of what is learned. Italso requires a particular kind of justification-scientificjustification based on integrative cross-disciplinary modelingand monitoring data-for changes in policies andimplementation measures. In principle, at least, it does notseem so very difficult a prospect to reduce those procedures andmodes of justification to a set of administrative law principlesaimed at providing the transparency, accountability, and"objective boundaries" that are currently lacking. To be sure,those legal principles will diverge from the familiaradministrative law principles that we have all come to knowand love, or in some cases to loathe. But that does seem amanageable task. One might even envision administrative lawproceeding on two tracks: ordinarily the familiar "fixed rule"track will apply, except in circumstances where the agency canjustify, according to well-understood standards, shifting to theadaptive management track, and at that point a second set ofadaptive management administrative law principles would kickin, requiring different procedures and further justifications forchanges in the course of action. It all sounds a bitbureaucratic, but perhaps that is the price we must pay to

44. Id. at 54.

2005]



19/21

MINN. J.L. SCI. & TECH.

reconcile science and law. Thoughtful, scientifically attunedlawyers like J.B. Ruhl will surely have a central role to play inworking this all out, but it strikes me as a valuable andexciting project.

Ruhl makes a further contribution, introducing theconcepts of "volatility" and "drift" to set some outer boundarieson back-end adjustments that would escape "obtrusive publicparticipation and judicial review." 45 This is a creative andnovel suggestion, worthy of careful consideration. My initialreaction, though, is a bit skeptical. Ruhl would confine earlychanges in position to a narrow range (to constrain volatility),gradually expanding the permissible range of volatility overtime, until an outer maximum range of permissible deviationfrom the initial position is reached (to constrain drift). Thisappears to leave room for much larger lurches in position asone proceeds outward in time from the starting point. It is notclear, however why large changes are more acceptable if theyoccur later in time, rather than early. In the beach mouse case,for example, suppose early monitoring results indicate that theinitial course of action is causing dramatic declines in thespecies, requiring early large-scale management adjustments toprevent ecological disaster. Those kinds of problems might justas plausibly show up early as late. On the other hand, ifmanagement is maintaining a steady course for a long period oftime, it might suggest that whatever management measuresare being pursued are actually working fairly well. From thatperspective, large deviations from a long-established andapparently successful course of management action look everybit as suspicious as earlier ones, and should be entitled to nomore deference.

On the other hand, ecological science has proven itselfcapable of generating very large surprises at any time.Sometimes we come to appreciate that our actions have beenpredicated upon very basic misunderstandings of ecosystemdynamics, necessitating large-scale corrections. There is noparticular reason to think those kinds of lurching advances inscience will come sooner, or later, in the process. Genuineadaptive management is almost certain to be somewhatvolatile, but unpredictably so as to when volatility will berequired, if we give it a chance to work. Nor is "drift" outsidepredetermined parameters necessarily a bad thing, if it is

45. See id. at 55-56.

[Vol. 7:1



20/21


predicated upon real advances in scientific understanding. Soit is a little hard to see just what is added by using the conceptsof "drift" and "volatility" as the basis for determining when toallow citizen participation and judicial review, however creativeand intuitively appealing the concepts may be.

At bottom, I suspect all this goes to the question of whatcounts as justification. The question is not how much do wewant to allow agencies to depart from an initial course ofaction, because we cannot presume that a smaller departure isthe sounder course of action, or that only small departures arejustifiable early but larger departures are necessarilyjustifiable later. The question is when do we want to allowsuch departures, of whatever magnitude. The answer, ingeneral, is when the departures are justified by advances inscientific understanding. That is what adaptive managementis all about, and that is the principle to which we must makeadministrative law adapt if we are to bring environmental lawand policy into congruence with contemporary ecologicalscience.

2005]



21/21

panarchy and adaptive change - around the loop and back again

Documents